Saliency Background Guided Network for Weakly-Supervised Semantic Segmentation

被引：0
作者：
Bai X. [1 ]
Li W. [1 ]
Wang W. [1 ,2 ]
机构：
[1] School of Computer and Information Technology, Shanxi University, Taiyuan
[2] Key Laboratory Computational Intelligence and Chinese Information Processing of Ministry of Education, Shanxi University, Taiyuan
来源：
Moshi Shibie yu Rengong Zhineng/Pattern Recognition and Artificial Intelligence | 2021年 / 34卷 / 09期
基金：
中国国家自然科学基金;
关键词：
Class Activation Response Map; Deep Neural Network; Image-Level Label; Saliency Detection; Weakly-Supervised Semantic Segmentation;
D O I：
10.16451/j.cnki.issn1003-6059.202109005
中图分类号：
学科分类号：
摘要：
Weakly-supervised semantic segmentation methods based on image-level annotation mostly rely on the initial response of class activation map to locate the segmented object region. However, the class activation map only focuses on the most discriminative area of the object, and the shortcomings exit, including small target area and blurred boundary. Therefore, the final segmentation result is incomplete. To overcome this problem, a saliency background guided network for weakly-supervised semantic segmentation is proposed. Firstly, the background seed region is generated through image saliency mapping and background iteration, and then it is fused with the class activation map generated by the classification network. Thus, effective pseudo pixel labels for training the semantic segmentation model are obtained. The segmentation process does not entirely depend on the most discriminative object. The information complementation is implemented through image saliency background features and class activation response map. Consequently, pixel labels are more accurate, and the performance of the segmentation network is improved. Experiments on PASCAL VOC 2012 dataset verify the effectiveness of the proposed method. Moreover, the proposed method makes a significant improvement in segmentation performance. © 2021, Science Press. All right reserved.
引用
收藏
页码：824 / 835
页数：11
相关论文
共 35 条
[11]  
VERNAZA P, CHANDRAKER M., Learning Random-Walk Label Propagation for Weakly-Supervised Semantic Segmentation, Proc of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 2953-2961, (2017)
[12]  
ZHU Y, ZHOU Y Z, XU H J, Et al., Learning Instance Activation Maps for Weakly Supervised Instance Segmentation, Proc of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, I, pp. 3111-3120, (2019)
[13]  
WAN F, LIU C, KE W, Et al., C-MIL: Continuation Multiple Instance Learning for Weakly Supervised Object Detection, Proc of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 2194-2203, (2019)
[14]  
PINHEIRO P O, COLLOBERT R., From Image-Level to Pixel-Level Labeling with Convolutional Networks, Proc of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 1713-1721, (2015)
[15]  
EVERINGHAM M, ESLAMI S M A, VAN GOOL L, Et al., The PASCAL Visual Object Classes Challenge: A Retrospective, International Journal of Computer Vision, 111, 1, pp. 95-136, (2015)
[16]  
ZHOU B L, KHOSL A, LAPEDRIZA A, Et al., Learning Deep Features for Discriminative Localization, Proc of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 2921-2929, (2016)
[17]  
KOLESNIKOV A, LAMPERT C H., Seed, Expand and Constrain: Three Principles for Weakly-Supervised Image Segmentation, Proc of the European Conference on Computer Vision, pp. 695-711, (2016)
[18]  
HUANG Z L, WANG X G, WANG J S, Et al., Weakly-Supervised Semantic Segmentation Network with Deep Seeded Region Growing, Proc of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 7014-7023, (2018)
[19]  
SUN G L, WANG W G, DAI J F, Et al., Mining Cross-Image Semantics for Weakly Supervised Semantic Segmentation, Proc of the European Conference on Computer Vision, pp. 347-365, (2020)
[20]  
PAN S Y, LU C Y, LEE S P, Et al., Weakly-Supervised Image Semantic Segmentation Using Graph Convolutional Networks, Proc of the IEEE International Conference on Multimedia and Expo, (2021)
1234